Intelligent computer systems rely on more or less complex computational entities that represent occurrences and events in the real world. Usually, such entities are formed from representational primitives called properties, attributes, features, etc. To reflect varying degrees of uncertainty, originating from human judgement and the intrinsic nature of the world, such property values occur as more or less vague linguistic symbols or exact numeric expressions. Determining similarity between two properties is usually done on either the symbolic or the numeric level. This seems to be too restrictive for case-based reasoning and similar approaches as these often face mixed specifications. In this paper we propose a flexible and systematic scheme for representing crisp properties and two types of fuzzy properties. It also provides a consistent mechanism to establish similarity scores for the various instance combinations.
|Number of pages||6|
|Journal||IJCAI International Joint Conference on Artificial Intelligence|
|Publication status||Published - 1997 Dec 1|
|Event||15th International Joint Conference on Artificial Intelligence, IJCAI 1997 - Nagoya, Aichi, Japan|
Duration: 1997 Aug 23 → 1997 Aug 29
ASJC Scopus subject areas
- Artificial Intelligence